In 2014, Dutch soccer club Heracles Almelo was in trouble. It hovered just above the relegation zone of the country’s top domestic soccer league, the Eredivisie. It desperately needed goals; a striker who could save it from the drop. But it had no money.
In days gone, the club would have sent out scouts and agents to secure uncertain talents. Instead it called Giels Brouwer, a young data analyst and soccer fanatic who had recently founded SciSports, an “intelligence-as-a-service” firm from the nearby city of Enschede.
Brouwer suggested Wout Weghorst, a lanky, 22-year-old forward with a modest reputation at second division team FC Emmen. Weghorst was not to be offered a new contract, and his career looked set to tank. But his stats said otherwise.
“He looked a little bit clumsy but scored a lot of goals,” Brouwer tells Red Herring. The ability to sign Weghorst for free sweetened the deal helped too. Within a year Weghorst was Heracles’ top scorer, and landed a spot in the Dutch under-21 national side.
In 2016 Heracles sold Weghorst to FC Twente, of Brouwer’s home town, for €1.5 million ($1.8m). He is now joint-second top-scorer in the Eredivisie and one of Europe’s brightest young talents.
SciSports is one of a growing number of data analytics companies that are changing the world’s richest and most popular sport. Since the early 2000s increasing numbers of clubs have forsworn traditional scouting and tactical methods in favor of Big Data and technology.
It has taken years to break down those barriers – and it’s still a fight for soccer’s data pros to convey their message to the sport’s elder statesmen. But they are succeeding. And soccer will never be the same.
Data has driven American sports for decades. It’s barely possible to imagine baseball and football without the floods of numbers and graphs that accompany every live cast. Cricket, the world’s second most popular sport, is a sucker for row upon row of dense, game-day information too.
But soccer has taken a lot longer to glom on. There are plenty of reasons. One is that it is, by nature, a fast, frenzied and seemingly free-flowing sport: one whose success experts have long placed on intangibles like spirit and individual brilliance.
That is still true. But increasingly, as the money in the game has exploded and teams look for any advantage they can, they have turned to teams of data specialists leveraging careers in betting and other number-driven verticals to make money in soccer.
Sam Allardyce, the former England manager, was an early data pioneer. During his time leading Bolton in the Premier League from 1999 to 2007, he used a tool called Prozone to consolidate his playing style, and to move effectively in the transfer market.
Jay Jay Okocha, Fernando Hierro and Youri Djorkaeff were some of the stars Allardyce brought to the small, northern English club, which, improbably, reached the UEFA Cup–now called the Europa League–Europe’s second most prestigious continental cup.
In 2007 Ted Knutson, a University of Oklahoma graduate, was working for PinnacleSports, a betting company. That year he became its lead trader for the English Premier League, the world’s most lucrative domestic market.
In 2014 Knutson joined Smartodds, a firm which was making headlines for its vanguard-leading data approach to soccer. Moneyball, a 2012 movie which charted the rise of Billy Beane’s data-led Oakland A’s baseball squad, accelerated interest in data.
Matthew Benham, a betting entrepreneur, bought Danish side FC Midtjylland a year later, vowing to bring Moneyball-style management to the ailing club. Within a year it had won the Danish Superliga–its first-ever trophy–and beat giant Manchester United during last season’s Europa League.
Benham also owns Brentford FC, which almost won promotion to the Premier League just a year after rising from England’s third tier, League One. “Moneyball changed the way people hired players,” says Knutson. “This kind of action doesn’t just start here – it goes on for generations.”
Brentford had the fifth lowest budget in the league but finished fifth, ninth and tenth of 24 sides in the years Knutson was there. Data is allowing clubs with smaller bankrolls to compete at wealthier tables, he says. “And then in the Premier League, the one thing we’re definitely certain of right now is that not everyone is taking advantage of all the different edges data could bring.”
Data like heat maps and possession statistics have become staples of the soccer viewing experience, and most halftime shows are flooded with pundits’ attempts to explain complex data points.
The most recent of these is “expected goals”, or “xG”, which measures how many goals a team should score based on the chances they create. It was created by Opta, which began collecting data on teams in 2009. Today its method involves three people per match: an analyst for the home team, the away team and somebody overseeing the process.
The analysts track every time the ball is touched in real time, recording the specific action. An average top-level game generates between 1,500 and 2,000 events. “This process allows us to capture every action in significant detail,” says OptaPro marketing coordinator Ryan Bahia.
Wenger himself attracted the ire of fans this month when evoking xG to explain why his players had fared well, in the wake of a 3-1 loss to Manchester City. “If you look at the expected goals, it was 0.7 for them and 0.6 for us,” he said. “It was a very tight game, they created very little, had very little number of shots on target, one more than us, that’s all.”
Arsenal had employed Chicago-based StatDNA from the 2011/12 season, paying it around $250,000 per year not to provide information to its league rivals. In 2014, buoyed by its results, longtime manager Arsene Wenger persuaded the club’s board to buy StatDNA for around $3.2m.
It was big news: not least that it showed the soccer data market to offer exits for those getting it right. Knutson agrees. But, he adds, it is an industry unlike others for one, vital reason.
“Your customer base is so different,” he says. “It’s B2B, essentially, but it’s also B2B that is very networked and connected. So getting introduced to the right people that matter is more challenging. It’s certainly not the same as marketing strategies and growth strategies that you’d use in other areas: it’s unique in this way.”
Bridging that gap between data and soccer’s elderly statespeople is one of the biggest challenges facing data firms. Some clubs have even suffered public spats between traditional scouting staff, and data analytics teams. Last month US data analyst Craig Kline left Championship club Fulham after a row over the side’s approach to recruitment.
“We’re already using some version of machine learning in our models,” says Knutson. “But if you get too esoteric, too detached from the way the football works, it’s difficult to explain to people. And what really matters is being able to execute the strategy in the boardroom and on the pitch.
“There’s a lot of natural skepticism when you’re talking to people who’ve been in the game 25 years,” Knutson adds. “And the reality is that we’re just conveying information.”
Brouwer, whose work is focused mainly in the Netherlands, Germany and England, has faced the same issue. “I compare it with the housing market,” he says. “You have the online platform where you can quickly filter which houses suit your needs. You will not buy a house purely based on the images you see online. But it makes you process it a lot easier.”
Data helps recruitment in a market that has suffered more than its share of abuse and subterfuge. Graft is a big problem throughout the game. Even now, Brouwer says, slow moving countries tend to be “the ones where other things are involved in transfers as well, like Italy, Turkey – those kinds of countries!”
But those who use data are changing the way the sport is played on a granular level. Coaches can now buy players who fit directly into their style of play, rather than taking a punt on somebody who may not be a good fit. That is amplifying tactics such as Jürgen Klopp’s ‘Gegenpressing’, or the famed ‘Tikitaka’ style adopted by now-Manchester City manager Pep Guardiola.
Bahia and his team have noticed some striking trends in recent years. “Using data at a basic level for example, you can identify your opponent’s playing style in possession, identifying whether they are patient in build-up, are shot-heavy, or favor a crossing approach,” he says.
xG has showed how unsuccessful long-range shooting is, for example – and coaches are adapting as a result. That might be bad news for fans who enjoy a 30-yard screamer. But it is good news for those who enjoy tight, close-knit passing styles, and the players who are good at them.
“Recruitment departments need to be very aware of the tactical requirements on the team,” says Knutson. “A Tony Pulis needs a totally different set of players to a Jürgen Klopp or a Guardiola…that makes a unique challenge, and the communication really matters.”
“It will become more of a game-theory game than it already is,” adds Brouwer.
Innovation counts too. Knutson has developed a product called Engine Room that can cross-reference a multitude of individual stats to give a ‘wagon wheel’ of data points. Brouwer is going in a different direction, developing 3D data from camera footage at matches.
Bahia, meanwhile, warns that clubs should not just look at data. It is one part of a huge and complex manpower problem, in a sport whose variables are infinitesimal.
“Data is just one piece of the puzzle in this instance,” he says. “How a player might settle in a new environment is just one example of what needs to be considered beyond data.”
Take Dutch forward Memphis Depay. Highly touted at PSV Eindhoven, in the Eredivisie, before a $33m move to Manchester United, Depay failed to fit into life in England and, by 2016 and aged just 21, was looking for a new challenge.
It was crucial to find a club that could accommodate Depay’s fast, direct technique, while providing him with the environment off the pitch to excel. He turned to SciSports. “We suggested four clubs to him, of which Olympique Lyon was one,” says Brouwer. “According to his playing style we thought he would fit there really nicely.”
He did. Since his move to France Depay has scored almost a goal every other game, and is once again regarded as one of Europe’s most talented wingers. Like Wout Weghorst, the old soccer world was keen to spit Depay out. Data gave them, and analytics experts, the chance to shine again.